STUDIES IN SPfODLE FOEMATION. 139

Studies in Spicule Formation.

VIII.—Some Observations on the Scleroblastic Development ofHexactinellid and other Siliceous Sponge Spicules.

IV. Woodland,The Zoological Laboratory, King's College, London.

With Plate 7.

THE contents of the present paper, though brief, representthe results of a year's careful inquiry into the manner inwhich the spicules of siliceous sponges are produced by thesilicoblasts of the organism. This subject haviug alreadyattracted so much, attention from spongologists, most of myresults necessarily have been to a large extent merely con-firmatory of those obtained by previous workers, aud these,of course, I shall mention as briefly as possible. A few ofmy results, on the other hand, are new, and in one or twoinstances corrective of former work.

My material has consisted of typical examples of bothsiliceous groups of sponges—the Triaxonida (Hexactinellida)and Tetraxonida (Tetractinellida and Monactinellida). Someof my specimens have been prepared by the osmic acid andpicro-carmine method, already described in previous Studies,and others (the majority) with borax-carmine. The formeris the better method. The Monactinellida (viz. young portionsor buds of Tethya lyncurium, Hymeniacidon san-guinea, Halichondria panicea, Esperella lingua,Siphonochalina coriacea, Axinella polypoides.

140 W. WOODLAND.

Suberi tes domuncula, Microciona atrasanguinea,Cliona sp. and Ectyon ovoides) and Tetractinellida(viz. young portions of Greodia gigas, Corticium cande-labrum, Thenea muricata, S te l le t ta carbonaria,and Chondril la nucula) were obtained partly fromNaples and partly from Plymouth; the Hexactinellida (viz.young portions or buds of Rossella podagrosa, Kpk.,E. antarctioa, Ctr., Anoxycalyx ijimai, Kpk., Anaulo-soma schulzei, Kpk., Hyalascus hodgsoni, Kpk., andAulorossella longstaffi, Kpk.) were in greater partsupplied to me from the British Museum (the " Discovery"material) at the suggestion of R. Kirkpatrick, Esq.,1 andby permission of Professor Sir E. Ray Laukester, andthe rest (viz. young portions of Euplectel la marshalli ,Acanthascus cactus, and Crateromorpha meyeri)•were sent to me at my request by Professor Ijima fromTokyo. I wish to thank Sir E. Ray Lankester, Mr. R.Kirkpatrick, and Professor Ijima for so kindly assistingme in this essential; I also wish to express my indebtednessto Professor Dendy for his ever-ready aid in connectionwith spicule nomenclature, for references to the literature ofthe subject, and for kindly reading over the manuscript ofthis paper.

HEXACTINELLIDA.

The scleroblastic formation of hexactinellid spicules hashitherto, so far as I know, been studied solely by Ijima (15,16) and by Schulze (31), and their results have been includedin three sumptuous monographs on hexactinellid sponges pub-lished in 1901 and 1904. Ijima has described and figuredseveral developmental stages of the hexasters and one or twoother types of hexactinellid spicules in Euplectella mar-shalli and Rhabdocalyptus capi l la tus ; Schulze de-scribed and figured the scleroplasmic investment of certain

1 Mr. Kirkpatrick has recently published a systematic description of thesesponges obtained by the National Antarctic Expedition (I).

STUDIES IN SP1CULE FORMATION. 341

spicules, which are adult, or nearly adult, in Tr ichas te r inaboreal is .

Notwithstanding the large quantity of material which hasbeen placed at my disposal and my thorough examination ofthis material, I have been very little more successful than thetwo investigators just named in discovering the early stagesin the development of hexactinellid spicules.1 The earlieststage which I have found is indeed somewhat younger thanthat figured by Ijima, but it throws very little, if any, extralight on the question as to the condition of the scleroplasmat the time of origin of the spicule and as to the earlymorphogenesis of the spicule itself. The young stageswhich I have figured were present in Eosse l la podagrosa,Kpk., the histology of which species was comparatively wellpreserved. I have only figured the scleroplasmic investmentof spicules from this and the allied species, R. an ta rc t i ca ,because in the other genera which I have studied the spiculeswere in precisely the same condition.

Ijima's remarks on the scleroplasm in connection with theyoung spicules described by him are as follows:—" A.hexaster begins its development as a hexactin. . . . Thehexradiate principals, during the entire period of developmentof both the floricome and the graphiocome, are imbedded ina body of protoplasmic substance, enclosing a crowdednumber of nuclei. This nucleated substance may not im-properly be called the scleroblast-mass. . . . At first, so longas the terminals are yet undeveloped or are very short, themass may be said to present a more or less octahedral shape,with somewhat concave surfaces and with rounded corners.In it the three axes of the principals are disposed similarly tothe axes in a crystal octahedron, the outer ends of the prin-cipals coming up vei'y close, but I think normally not quite,to the surface at the six rounded corners. The mass mayotherwise be described as having its surface raised into six,radially directed, hump-like protuberances by the six prin-

1 It is worth remarking that the buds of my hexactinellids contained feweryoung spicules than the older portions of the sponge body.

142 W. WOODLAND.

cipals contained within. Later, after the terminals haveconsiderably advauced in growth, the scleroblast-mass appearsat the centre of the developing hexasters as a more sphericalbody, not unlike a berry, on account of the aggregatednuclei." "Not a trace of cell-outlines is discernible aroundthem [the aggregated nuclei], which fact makes me believethat the scleroblast-mass represents a syncytium." And,referring to the prevalent but mistaken notion that thescleroblast nuclei change their position in order to superintendthe process of deposition in different regions of the spicule,Ijima remarks :—" Certain it seems that, during the growthof the terminals, no nucleus moves away from its grouparound the spicular centre. At least I could gather noevidence pointing to such a movement. I t is true that, aftera certain period in the growth of terminals, a variable numberof nuclei is met witli riglit away, or in close proximity to,these. However, they are altogether so inconstant in numberand indefinite in position that it is exceedingly questionableif they have anything to do with the building up of theterminals." My figures confirm all the preceding remarks.Fig. la depicts a stage in which the terminals have not nearlyreached the surface of the spherical syncytium, and which is,therefore, younger than any figured by Ijima. Ijima goes onto remark concerning the " great probability . . . that eachgrowing terminal is completely invested by an extremely thinprotoplasmic layer, specialised physiologically at least as thesecretive matrix, and standing in direct continuity with thescleroblast-mass. Such a layer, however, could never beclearly demonstrated."1 I may say that I have many timesobserved in my preparations the film of scleroplasm whichinvests the terminals of both young and old spicules (andwhich I have endeavoured to represent in figs. 3 and 6),although this is not by acy means always visible; in fact,usually is not.

As shown by my figures the scleroblast-mass varies con-1 Ijima, however, in a later contribution (16) describes and figures tbis

layer in the oxyhexasters of Rhabdocalyptus capillatus.

STUDIES IN SPICULE FORMATION. 143

siderably in size, and this feature is probably related to thedifferent kinds of spicules to be produced. So few reallyyoung stages exist in my material that I am quite unable tosay of what types of spicule my figures repi-esent the youngstages ; possibly it would be difficult to say under any circum-stances, since the young hexact must be an initial develop-mental stage common to all types, both megascleres andinicroscleres.

Further, although I am unable to demonstrate it by actualexamples,1 I may say that I think it is highly probable thatthe initial granule from which in all probability thesix rays of the hexactin grow out, is found within asyncytium, and not, as all other siliceous spongespicules are, in a single cell. In support of this state-ment I may adduce the following evidence : (a) cell-clusterssimilar to those represented in fig. 1, are occasionally foundin the tissues of the sponge, and are apparently identicalwith those enveloping young spicules; (I) the large numberof nuclei present in the syncytium containing the youngspicules figured confirms the view that more than one nucleuswas present at the origin of the spicule; and (c) this distinc-tion between the modes of origin of hexactinellid and othersiliceous sponge spicules is only one of many distinctionswhich separate the Hexactinellida as a group from othersiliceous sponges.

The great extension of the rays of the megasclere type ofspicule during the later stages of development in every casecauses the centrally-situated syncytial mass to decrease insize and in some cases to vanish, since the scleroplasm isneedad for peripheral growth. Ijima suggests that, siuce incalcareous sponge spicules the scleroblasts ultimately desertthe spicule, the syncybium of the hexactinellid spicule pos-sibly does the same, and, as is well known, numerous otherauthors have often asserted that this desertion does actually

1 1 searched with the greatest care for a syncytium containing a siliceousgranule, spherical or six-cornered, but, as stated in the text, I met with nowell-defined stage younger than that figured.

144 W. WOODLAND.

occur in the case of other sib'ceous sponge spicules. Forseveral reasons I very much doubt if this assumption of thedesertion of the siliceous spicule, hexactinellid or otherwise,by the scleroplasm, is justifiable, at least in the majority ofcases. It must be remembered that the scleroplasm whichbuilds up the triradiate and other spicules of calcareoussponges differs from the syncytial scleroplasm of siliceousand, indeed, mauy calcareous spicules, in that it consists ofseparate scleroblasts, cylindrical or semi-cylindrical in shape(5), which are not united to form a syncytium entirelyenveloping the spicule, and which are in consequence free toleave the spicule at almost any period of deposition. On theother hand, seeing that the syncytium which envelops acomplex hexactiuellid inegasclere, e.g. (and the rnicro-sc le res certainly retain their syncytia) is correspondinglydistended and configurated, it is difficult to understand howthe syncytium could take its leave as a whole, and it isexceedingly doubtful if the small portions containing nuclearmatter separate themselves off from the rest. Personally Ibelieve the scleroplasm of all or, at any rate, most siliceoussponge spicules clings to the spicule as long as it persists,although in many cases it becomes so attenuated as to beinvisible. Indeed, the presence of a nucleus is frequentlythe only visible indication of the existence of a scleroplasmicfilm enveloping the spicule.

Ijirna again suggests the probability that the six rays ofthe hexactinellid spicule originate as separate sclerites, butsupplies no evidence in support of this view, merely men-tioning the mode of origin of the triradiates and quadri-radiates of Calcarea as possibly presenting an analogouscase. That no analogy can be drawn from the mode ofdevelopment of these calcareous spicules is, I think, suffi-ciently manifest on remembering that each of the threemonaxous of the calcareous triradiate is produced from, andthroughout subsequent growth is constantly related to, oneof the three cells which " fused " to form the " trefoil" (4, 5),a feature entirely absent in the case of the hexactinellid

STUDIES IN SPIOULE FORMATION. 145

spicule, in which no "cells" exist, and in which the nucleiof the undoubted syncytium bear no definite relation to theconstituent rays. Moreover, as far as I know, there is noauthenticated instance of any siliceous spicule arising by thefusion of initially-separate parts contained in an individualsyncytium, though of course it may well be that the hexacti-nellid spicules are peculiar in this respect as in others. Theonly ground for believing that the hexact spicule arises bythe central fusion of six initially-separate rods that I canimagine is the general similarity which many hexactinellidspicules bear to certain Radiolarian skeletons. In Radio-larians these rays are usually separate, at any rate initially,and in some cases they become fused centrally. If the modeof deposition of hexactinellid rays in the spherical syncytiumcan be shown to be at all similar to that of Radiolarianspines in the spherical body, a reason certainly exists forsupposing that the hexact rays may be deposited initially ina separate condition, but there is no other reason for thissupposition that I know of, and personally I believe the sixrays of the hexact spicule will be found to emerge from aspherical granule.

Tet rac t ine l l ida and Monactinell ida.

Generalising the large number of statements1 which havebeen made by the various authors enumerated in the biblio-

1 To avoid burdening the text with a somewhat unnecessary account ofprevious work on the scleroblastic development of siliceous sponge spicules—an account of which both Minchin (4) and Maas (24) have provided an outline—I append a full bibliography of the subject, mentioning after each paper thetype or types of spicule, the scleroblastic developm'ent of which is describedwithin. I may here add that I have not been able to trace the developmentof every type of spicule contained in the genera mentioned in the Introduc-tion, and indeed only in one or two cases would this have been desirable (e. g.the development of the rosettes of Esperella), since the development of mosttypes is already known. In fact, so far as I am aware, only in Cbondrillanucula (Tetractinellida), and in Siphonochalin a coriacea, Axinellapolypoides, Suberites domuncula, Microciona atrasanguinea,Cliona sp., and Ectyon ovoides (Monactinellida), among the sponges

146 W. WOODLAND.

grapky concerning the scleroblastic development of: tetra-xonid and rnonaxonid spicules, we may say without hesita-tion that the vast majority of the spicules, widely as theymay differ among themselves with respect to form and size,arise each as a granule contained within a single scleroblast.The few siliceous spicules which do not thus arise from asingle centre, but from several granules contained within thescleroblast (e.g. orthodragmata, Carter's bundles of " tricur-vates," etc.) are exceptional, and in every one of thesecases the " spicule" consists of several parts throughout itsexistence. I am not aware of a single instance in whichthe body of a siliceous spicule-individual is built up by thefusion of originally separate parts, i. e. visible parts, thoughstatements have recently been made to the contrary (seebelow).

Further, with the exception of those spicules which attaina relatively large size (e. g. the monaxons of Bsperella,Ectyou, and Tethya, the large tetracts of Geodia and thespherasters of Tethya) all these spicules throughout theirexistence as skeletal elements of the sponge-body remainenveloped by the distended substance of the s ingle sclero-blast in which they arose; in other words, the nucleus of thescleroblast does not divide (figs. 10—18).

I t is needless for me to enter into details in connectionwith the morphogenesis of each type of spicule; it suffices tosay that careful work has proved that in the vast majority ofcases each spicule, whatever its ultimate shape, originates asa granule, and only by degrees assumes its final form whilstenveloped by the formative scleroplasm. Also, so far as weknow, all growth of the spicule is accretionary and neverinterstitial, and this fact is constantly to be borne in mindwhen the identification of young forms of certain types ofspicule is in question.

As stated above, the building up of siliceous spongespicule-individuals by the secondary fusion of separate de-I have studied, has the scleroplasm associated with the spicules not beenhitherto described (see figs. 10—17).

STUDIES IN SPICULE FORMATION. 147

posits is unknown; in other words, no one has yet foundsiliceous equivalents of the triradiates and quadriradiates ofCalcarea. This being the case, it was with some interestthat I read an account of this very process which was statedto occur in connection with the spicules of Te thya l yncu -rium. This account was contained in a paper published byDr. 0. Maas in 1900 (24). The exceptional nature of thestatements in this paper led me to pay special attention tothe scleroblastic development of the spicules in T e t h y alyncur ium, but, as on a former occasion,1 I was unable toconfirm Dr. Maas' results. Maas makes the following state-

1 To enable the reader to estimate the value of Dr. Maas' reply (3) I will recallthe facts. Minchin in 1898 (4) showed that each triradiate spicule of the Asconsponges which lie investigated originated as three separate rods radiatingfrom a common centre, each ray being produced by a pair of cells of the" sextet," and that these rods only secondarily became fused together at theirinner extremities to form the triradiate spicule. In the same year ( la) andin 1900 (2) Maas published papers, each dealing in part with the similarspicules in Sycons. In the latter paper (and with this 1 am chiefly concerned)Maas described a process of spicule formation as occurring in Sycons which isradically different from that which occurs in Ascons—each triradialc spicule,as a whole, being stated to arise as a single concretion in one mother-cell. AtProfessor Minchin's suggestion I, during 1903 and 1904, devoted the greaterpart of a year's work to ascertaining for certain whether the process of spiculeformation in Sycons was so fundamentally different from that obtaining inAscons, as alleged by Maas, and.I found, both to Professor Minchin's and myown satisfaction, that the mode of spicule formation in three species of Syconclosely allied to those studied by Maas was identical with that found inAscons. This being the case I considered that Maas' statements were erro-neous, and I ventured to express that opinion—perhaps in too brusque amanner. Dr. Maas in his counter-criticisms says that I obtained anomalousresults owing to the method of fixing the Sycons which I adopted—a criticismwhich applies to Professor Minchin's work as well as my own. I can onlyreply that Professor Miiicliiu's slides (and my own to a less extent) have beeninspected by numerous zoologists, and that no fault has ever been found withthem on the score of imperfect histological preservation. Maas' criticismsimply proves, what indeed he admits, that he has never seen the three-rodstage of the triradiate spicule, else he would not assume it to be an artefact.Maas again speaks of the initial stages of spicule deposition as if, owing, as hestates, to their small size aud transitoriness, it were almost impossible loobserve them. This is not the case. They are easily to be observed, pro-

148 W. WOODLAND.

merits: (a) the tylostyle spicule arises by the fusion ofseparate granules present in the substance of the mother-scleroblast; (b) the chiaster spicule arises by the junction atthuir inner extremities of several rods of silica which, aredeposited in the sclei'oblast so as to lie radially in the cell-sphere ; and (c) the spheraster usually arises by the fusion oftwo or more calthrops (each of which arose as a granule con-tained by a single scleroblast), which serve as a basis onwhich the rest of the silica is deposited—from which it isinferred that Tethya has tetractinellid affinities. This laststatement is illustrated by a somewhat extraordinary figure(his fig. 33) of two closely-apposed attenuated calthropsspicules, each contained in its scleroblast. It is significantthat if these three statements be true, then, as before stated,spicule formation in T e t h y a l yncu r ium proceeds in amanner different from that hitherto found in any othermonactinellid, or, for that matter, tetraxonid sponge.

With regard to the first statement, I entirely agree withWeltner's estimate of its value (38). Weltner rightly con-siders that the several granules have nothing to do with thetylostyle spicule, and that this arises, like other siliceousmonaxons, from a single granule which gradually elongatesand assumes the form of the adult spicule by continueddeposition of silica on its surface. Further Maas' figures of

vided that plenty of suitable material, i.e. a quantity of very young Sycons, isto hand, that this material is prepared in the way described by Minchin, andthat tbe observer possesses sufficient patience to find the comparatively fewyoung stages in his preparations. Maas' remarks about what he supposes tobe my ignorance of his own paper and other relevant literature are incompre-hensible to me siuce, besides being entirely mistaken, they are entirely irre-levant to the subject under discussion. Because, as Maas points out, Iadopted Mincliin's convenient summary ('Zoological Record* for 1901) ofhis statements concerning spicule formation (and Maas does not dispute tbeaccuracy of Mincbin's rendering of bis statements) I fail to see that anyoneis justified in drawing the curious conclusion that I was unacquaintedwith the paper which I was criticising! I may add that in my previouspaper (5) I inadvertently ascribed "tuning-fork" spicules to Clathrinalacunosa.

STUDIES IN SPIOULE FORMATION. 149

the cells in connection with the adult tylostyle are quiteincorrect. These never have either an epithelial1 dispositionor the " syncytial" disposition shown; on the contrary, eachof the several " c e l l s " (i.e. nucleated portions of the syn-cytium) in connection with the elongated spicule forms asloping mound-like mass containing the nucleus, and lies inclose apposition with the spicule, as depicted in my figure ofthe Esperella monaxon (fig. 19).

Maas' second statement I believe to be not more validthan the first. The chiaster, according to my observations,originates, like the spheraster, as a granule contained withinthe cell, and this granule, also like that of the spheraster,develops rays on its surface (fewer in number2 and relativelylonger and more uniform in thickness as compared with thoseof the spheraster) which become those of the adult chiasfcer(figs. 20 c, g). The mother-cell, as Maas states, remainsundivided.

Maas' third statement is also contradicted by my observa-tions. I can find no evidence whatever that the spherasterever passes through a calthrops stage; on the contrary, itdevelops in almost exactly the same way as the chiaster.The initial grannie develops numerous radiating rays whilstquite small (Maas' fig. 7 is quite correct e. g.), and from thisground form (figs. 20 b, d) the adult spicule is formed(figs. 20 h, j , h) by the further accretion of silica, and also,I believe, by the occasional development of fresh rays duringthe early stages. As implied above, these rays of the sphe-raster are more numerous and more conical in form thanthose of the chiaster, and these differences alone distinguishthe younger developmental stages; at first indeed it is very

1 Maas makes the same mistake in connection with the monaxon spicule ofSjcandra, figuring the half dozen scleroblasts as spherical cells justtouching the surface of the spicule. Scleroblasts never have this arrangementin connection with spicules or, so far as I know, elsewhere.

2 Even if, as occasionally may be the case, these spines of the developingchiaster are four in number, this fact affords no justification for Maas' state-ments in connection with the spheraster.

150 W. WOODLAND.

difficult to distinguish between them. The development ofthe Tethya spheraster in fact is essentially the same as thatdescribed by Keller1 (18) for the spheraster of Chondr i l l anucula . I may further mention that Miss Sollas (36) alsofailed to find any signs of calthrops, though plenty ofglobules, in the larva of T e t h y a inga l l i . The singlescleroblast which contains the initial granule divides as thespheraster increases in size, with the result that the adultspheraster possesses several cells; whether any of these arenot division-products of the original mother-cell but areextraneous in origin, as Maas confidently thinks, I am unableto say, and I fail to see how it is possible to say without atleast very special inquiry.

With the above criticism on Maas' woi'k in connection withspicule formation I may include another and minor one.During my observations of tbe developing spicules ofEspe re l l a l ingua I could never find more than one nucleusassociated with each of the anisochelee in this species (fig. 18)whereas Maas (23) states that four nuclei occur in connectionwith each such spicule in E s p e r e l l a lorenzi—a statementwhich I feel sure is incorrect. I have observed numerousrosettes in E. l ingua , and in every case each anisochele onlypossessed one nucleus. In each rosette the anisochelaewereembedded at their inner extremities in a spherical mass ofpi-otoplasm wliich contained several nuclei, but all of thesenuclei were quite distinct from those which belonged to theanisochelas. The answer to the question as to how the aniso-chele rosette is formed is not as yet known. Carter (8), withothers, supposes that the entire cluster of anisochelfe isderived from separate silica deposits initially present in one

1 Or rather as Keller should have described it. Keller figures the spikesas arising successively on the surface of the siliceous globule, whereas, inreality, they arise for the most part simultaneously as small knobs whichgradually assume the pointed form. The spicule is enveloped in one sclero-blast throughout its existence. The calthrops of Corticium candelabrume.g. also originates in the same way—the granule developing four processeswhich elongate to form the four rays of the calthrops; also the chiaster,oxyaster, and sterraster of Geodia -g'gas.

STUDIES IN SPICULE FORMATION. 151

cell, and even states that he " observed in one cell twoinequianchorates together end to end or slightly overlappingeach other; but this was in the equianchorate stage."Whether this interpretation of the origin of the rosette is oris not the correct one future inquiry must determine. It isquite certain that the trichites of the ortho- or trichodragmataare produced in one cell and remain in one cell throughouttheir existence, i. e. the nucleus never divides. Carter alsodescribes and figures the production of several " tricurvates "(toxas) in one cell, the nucleus of which similarly neverdivides, and several other kinds of dragmata are known. Ithink it highly probable that subsequent inquiry will provethe Esperia rosette to be derived from a cell-clustei*, each ofthe peripheral cells of which produces an anisochele, thecentral cells remaining as the central multinncleated " spheri-cal mass of protoplasm " described above.

In conclusion I may remark upon the significant fact thatalthough all sponge spicules arise in the interior of cells, yetin each of the three great groups of sponges—the Tetraxonida,Triaxonida, and Calcarea—spicule formation proceeds onvery different lines. In the Tetraxonida the spicule typicallyarises and continues to exist in one cell, i. e. the cell entirelyenvelops the spicule; in some cases the nucleus of this celldivides so that a syncytium is formed, and in a few instancesseveral " spicules " are produced in the interior of a singlecell. In Triaxonida all the evidence points to the conclusionthat a large spherical syncytium containing many nuclei mustbe formed before the spicule is secreted in its interior, andthat the three-dimensioned spicnle (the six rays of which ariseas outgrowths from an initial granule) is situated in this syncy-tium, at least for some period of its growth, in much the sameway that certain Radiolarian spines are enclosed in the sphe-rical Radiolarian body. Finally in Calcarea, at least themajority of spicules (i. e. with the possible exception of cer-tain monaxons) are formed by the apposition of cells in twos(either by cell-division or cell-junction) and threes (solely bythe junction of separate cells) which do not entirely envelop

152 W. WOODLAND.

the spicule as in the other two groups, but merely adhere tothe spicule-ray as cylinders or semi-cylinders. These differ-ences as regards the method of spicule formation in thethree groups of sponges doubtless possess phylogeneticsignificance.

SUMMARY.

1. The earliest stages in development of the Hexactinellidspicule are at present unknown, but there exist reasons forsupposing that it originates as a granule enclosed by aspherical syncytium in which cell-outlines are absent, andthat the six rays of the hexact grow out from this granule.

2. The earliest stage of development yet discovered is thesmall hexact, the rays of which do not extend to the peripheryof the enveloping syncytium.

3. The rays of the hexact elongate, causing the sphericalsyncytium at first " to present a more or less octahedral shape,with somewhat concave surfaces and with rounded corners "(Ijima).

4. The rays at length extend beyond the spherical contourof the syncytium (the scleroblast mass), the scleroplasm ofwhich, however, persistently adheres to the rays as a thinfilm, which occasionally includes nuclei.

5. The peripheral growth of the megasclere type of spiculecauses the spherical syncytium enveloping the point of junc-tion of the six rays to dwindle and finally to disappear onaccount of the distension involved. The whole of the micro-sclere, with the exception of the terminals, remains per-manently enveloped by the spherical syncytium.

6. All tetractinellid and monactinellid spicules originate asgranules contained within single cells. In a few instancesthe spicule arises from several granules within the cell andthen consists of separate parts (dragmata, for example).

7. All growth is accretionary. There is no well-authenti-cated instance of a siliceous sponge-spicule being formed by

STUDIES IN SPIOULE FORMATION. 153

the fusion of at-first-separate parts—such as occurs in cal-careous sponges, for example.

8. With the exception of very large spicules, the nucleusof the distended scleroblast remains single throughoutgrowth.

9. It is a significant fact that spicule formation proceedson very different lines in each of the three great groups ofsponges—the Tetraxonida, Triaxonida, and Calcarea.

Note.—Unless exceptional opportunities for further workin connection with spicule formation should present themselvesthat it is intended the present Study shall conclude the series.

LITERATURE.

A. PAFEKS NOT CONCERNED WITH THE DEVELOPMENT OF SILICEOUS

SPICULES.

1. KIRKPAXKICK, 11.—"Hexaotinellida: National Antarctic Expedition,"vol. iii, 1907, B.M. (N.H.).

1 a. MAAS, 0.—" Die Ausbildung des Kanalsystems uud Kalkskeletts beijungen Syconen," ' Verb. Deutscb. Zool. Ges.,' viii Jalir., 1898.

2. "Die Weiterentwickluug der Syconen nach der Metamorphose,"•Zeitscbr. f. wiss. Zool.,1 Bd. lxvii, 1900.

Ueber die Einwiikung Karbonatfreier uud Kalkfreier Salzlos-ungen auf erwachsene Kalkschwamme und auf Entwicklungsstadienderselben," ' Archiv Entwicklung. Org.,' Bd. xxii (4), 1906.

4. MiNCHiN, E. A.—"Materials for a Monograph of the Ascons: I," 'Quart.Journ. Mic. Sci.,' N.S., vol. xl, 1898.

5. WOODLAND, W.—"Studies in Spicule Formation: I. Sycon Spicules,"' Quart. Journ. Micr. Sci.,' N.S., vol. xlix, 1905.

B. PAPEKS CONCERNED WITH THE DEVELOPMENT OP SILICEOUS SPICULES.

8. CARTKB,, H. J.—"A Descriptive Account of the Fresh-water Sponges inthe Island of Bombay, etc.," 'Ann. Mag. Nat. Hist.' (2), vol. iv, 1849,p. 81 (the opinion is expressed that the spicules originate in tbe

7. "On tbe Ultimate Structure of Spongilla, etc.," 'Ann. Mag. Nat.Hist.' (2), vol. xx, 1857) p. 21 (the spicules of Spongilla are recognisedas arising each in one cell).

VOL. 52, PART 1.—NEW SERIES, 11

154 W. WOODLAND.

8. CAUTEB,, H. J.—"On the Nature of the Seed-like Body of Spongilla;on the Mother-cell of the Spicule, etc.," 'Ann. Mag. Nat. Hist.' (4),vol. xiv, 1874, p. 97 (development of the sigmata, anisochelre and toxas(several in one cell) of Esperella oegagropila).

9. "Development of Marine Sponges, etc.," 'Ann. Mag. Nat. Hist.'(4), vol. xiv, 1874, p. 321 (development of toxas of Esperella

gagropila and Microciona arraata).

10. " Further Instances of the Sponge Spicule in its Mother-cell,"'Ann. Mag. Nat. Hist.' (4), vol. xiv, p. 456 (toxas of Esperellaojgagropila and Microciona armata in one cell).

11. DELAGE, Y.—" Embryogenie des Sponges, etc.," 'Arch. Zool. exp. etg6n,' (2), vol. x, 1892 (monaxons of Spongilla and Reniera densa

arise in one cell).

DENDY, A. (see RIDLEY AND DENDY).

12. EVANS, K.—"The Structure and Metamorphosis of the Larva of Spon-gilla lacustr is ," 'Quart. Journ. Micr. Sci.,' N.S., vol. xlii, 1899(monaxon of Spongilla arises in one cell).

13. "A Description of Bphydatia blembingia, etc.,' 'Quart.Journ. Micr. Sci.,' N.S., vol.xliv, 1900 (development of Amphidiscs ofEphydatia in one cell).

14. GOETTE, A.—"Untersuchungen zur Entwicklungsgeschichte von Spon-gilla fluviatilis," Hamburg and Leipzig, 1886 (development ofamphidisc in one cell).

15. IJIMA, I.—"Studies on Hexactinellida: Contribution I," 'Journ. Coll.Japan,' vol. xv, 1901 (development of certain hexasters described inpart).

16. "Studies on Hexactinellida: Contribution IV," 'Journ. Coll.Japan,' vol. xviii, 1904 (further examples of scleroplasm associatedwith hexasters).

17. KELLER, C.—"Studien uber Organisation und Entwicklung der Chali-neen," 'Zeitsch. wiss. Zool.,' vol. xxxiii, 1879 (development of spiculesof Chalinula fertilis).

18. "Die Spongienfauna des rothen Meeres," ' Zeitschr. f. wiss.Zool.,' Bd. lii, 1891 (development of the spherasters of Chondrilla andmonaxons and sterrasters of Placospongia).

19. KOLLIKEE, A.—"Icones Histologicte," Abth. i, Leipzig, 1864 (monaxonand amphidisc of Spongilla described and figured in single cells).

20. LENDENEELD, R. VON.—"Die Tetractinelliden der Adria, etc.," 'Denk.Akad. wiss. Wien, Math, naturw. Classe,' vol. lxi, 1894 (statementconcerning asters of Geodia and Ancorina),

STUDIES IN SPICUL15 FORMATION. 155

21. LTEBEEKUHN, N.—"Beitrage zur Entwicklungsgeschichte der Spongillen,'''Arcli. Anat. Pliysiol. J. Miiller,' Hefts iv, v, 1856 (monaxons ofSpongilla arise each in one cell).

22. MAAS, O.—" Ueber die Entwicklung des Siisswasserschwamines," 'Zeifc.wiss. Zool.,' Bd. 1, 1890 (monaxon of Spongilla in one cell).

23. "Die Metamorphose vou Esperia lorenzi, etc.," 'Mitl.h. Zool.Stat. Neapel,' Bd. x, 1892 (the cells associated with the anisochelre ofthe rosettes described—probably incorrectly).

24. i! Ueber Entstehung und Wachstum der Kieselgebilde bei Spon-gien," 'SB. Akad. wiss. Munch.,' Bd. xxx, 1900 (development ofspherasters, chiasters, and tylostyles of Tethya lyncurium incor-reetly described).

25. RIDLEY, S. O., ANDDENDY, A.—" Challenger " Report, vol. xx, "Mona-xonida," 1887 (monaxonid spicules described as originating each inone cell; toxa ofEsperel lasimonis and anisochela of Cladorhizainversa so figured).

26. SCHMIDT, 0.—Spongien des adriatischen Meeres," I Supplement, Leipzig-,1S64 (development of spicules of Reniera sp.).

27. ' Zool. Ergeb. d. Nordenfahrt,' 1872 (chelae, sigmas, and ortho-dragmas of Esperella lucifera described as arising each in onecell).

28. SCHULZE, P. E.—" Untersuchungen iiber den Bau und die Entwicklungder Spongien: die Pamilie der Chondrosidse," ' Zeitschr. wiss. Zool.,'Bd. xxix, 1877 (stellate spicules of Chondrosia).

29. "Unter,,etc.: NeunteMitt; diePlakiniden,"'Zeifc. wiss. Zool.,'Bd. xxxiv, 1880 (monaxon and calthrops of Plakina described asoriginating in one cell).

30. "Unter., etc.: Zehnte Mitt; Corticium candelabrum, O.Sch.," ' Zeit. wiss. Zool.,' Bd. xxxv, 1881 (calthrops of Corticum inone cell).

31. " Wissenschaftliche Ergebnisse der DeutschenTiefsee-expeditionauf dem Dampfer 'Valdivia,' 1898-1899: Hexactinellida," Bd. iv,1904 (scleroplasm in connection with hexasters of Trichasterinaborealis).

32. SOLLAS, W. J.—"The Sponge Fauna of Norway, etc.," 'Ann. Mag. Nat.Hist.' (5), vol. v, 1SS0, pp. 130, 241, 396, 401 (one-cell origin of thetrichites ofSte l le t tanormani , the sterrasters of Geodia barre t tand the spherasters of Isops phlegrtei).

156 W. WOODLAND.

33. SOLLAS, W. J.—"The Sponge Fauna of Norway, etc.," 'Ann. Mag. Nat.Hist.' (5), vol. ix, 1882, pp. 141, 426 (one-cell origin of the sterrastersof Pachyinatisma johnstonia, the oxeotes and stigmata ofTetilla cranii, and the oxeotes, spherasters, grapnels, and otherspicules of Thenea wallichii).

34. "Challenger" Report, vol. xxv, "Tetractinellida," 1888 (one-cell origin of many spicules of the Choristida, the calthrops andrhabdus of ihe Lithistid desma, the spherasters of Tethya lyucu-riuin, and the trisens of Thrombus challengerii).

35. 'Encyclopaedia Britnnnica,' 9th ed., Article "Sponges," 1891(one-cell origin of many of the spicules named in the four precedingreferences).

36. SOLLAS, I. B. J.—"On the Sponges collected during the 'Skeat' Expe-dition to the Malay Peninsula, 1899-1900," ' Proc. Zool. Soc,' vol. ii,1902 (siliceous globules in Tethya laiva).

37. VOSMAEK AND PEKBLHAKING.—"Observations on Sponges/' ' Verh. Ak.Amsterdam' (2), Bd. vi, 1898 (morphogenesis of the anisochela ofEsperella syrinx).

38. WEI/TNEKJW.—"Siisswasserspongien von Celebes (Spongillidenstudien),"'Archiv f. Naturgeschichte,' Bd. lxvii, Beiheft, 1901 (Maas' accountof the origin of the tylostyles of Tethya criticised).

EXPLANATION OP PLATE 1,

Illustrating Mr. W. Woodland's " Studies in SpiculeFormation." (VIII.)

HEXACTINELLIDA.

Figs. 3, 5, and 9 are from preparations of Rossella antarctiea, Ctr.; theremainder (figs. 1, 2, 4, 6, 7, 8) are from R. podagrosa, Kpk.

FIG. 1 (X 1600 diam.).—Young hexact stages of development. 1 a repre-sents a hexact small in comparison with the enveloping syncytium. Thepink hue uf the spherical mass of scleroplasm is due to the presence ofnumerous nuclei which, owing to imperfect preservation of the tissues, arenot always visible individually.

FIG. 2 (x 1600 diam.).—Microscleric monodiscohexaster with somewhatdilated extremities. It is not certain whether the extremities bave yetassumed their final form, but it is probable,

STUDIES IN SPICULE FOEMATION. 157

FIG. 3 (x 2600 diam.).—Adult microscleric holoxyhexaster. Two nucleihave migrated from the ceutral spherical mass of scleroplasm.

FIG. 4 (X 1600 diam.).— Holoxyhexaster, about half-growu.FIG. 5 ( X 1250 diam.).—Adult microdiscohexaster, with several peri-

pherally-situated nuclei ("cells"). Bach terminal is invested with a thinfilm of scleroplasm, though this is not always visible.

FlG. 6 (X 2000 diam.).—Half-grown and adult forms of another variety ofmicrodiscohexaster.

F ie 7 (X 1600 diam.).—Central portion of an oxydiactine, showing theenveloping scleroplasm with distributed nuclei. The position of (.lie fouraborted rays is indicated by the mid-way swellings and axial cross.

FIGS. 8 (x 700 diam.), 9 (x 625 diam.).—Hexactines with the enve-loping scleroplasm well shown.

MONACTINELLIDA.

FIG. 10 (x 800 diam.).—Tylostyle of Microciona atrasanguinea inits scleroblast.

FIG. 11 (x 1600 diam.).—Acanthotylostyle of M. atrasanguinea in itsscleroblast.

FIG. 12 (x 800. diam.).—Tylostyle of Suber i tes domuncula in itsscleroblast.

FIG. 13 ( X 1600 diam.).—Oxeote of Siphonochalina coriacea in itsscleroblast.

FIGS. 14, 15 (x 1600 diam.).—Stylus and oxeote of Axinella poly-poid es in their scleroblasts.

FIGS. 16, 17 (x 1600 diam.).—Acanthoxeote and spiraster of Cliona sp.in their scleroblasts.

FIG. 18 (X 800 diam.).—Anisochela (one of a rosette) of Espcrellalingua in its scleroblast.

FIG. 19 (x 800 diam.).—Portion of tylostyle of E. lingua showing two ofthe several nuclei present iu the syncytium enveloping the spicule.

FIG. 20.—Development of the chiaster and spheraster of Tethya lyn-curium. a and e represent the first appearance of the spicule as a siliceousglobule. The rays appear on the surface of these globules much earlier insome cases than in others; cf. e.g. figs, b and e. As stated in the text, it isimpossible to distinguish between the very young stages of the chiaster andthe spheraster. A comparison of fig. g with figs, h and./ shows well, on theother hand, that it is quite easy to distinguish adult cliiasters from youngspherasters of the same size.

VOf,. 52, PART 1. NEW SERIKS. 12

VU.SZXS 3*1.1

3. n

S I L I C E O U S S P O N G E S P I C U L E S .